[gmx-users] How to simulate a protein that contains two cuban cluster

[Mark Abraham]

Stefan Schöbel wrote:
> Hei, 
> 
> I want to simulate a  protein that contains two cuban clusters with gromacs, so I want to know which forcefield is recommended to use for Fe-S compounds.

Probably, none of them. One has to develop parameters for these atom 
types based on experimental or high-level computational data. The odds 
against data on compounds even vaguely resembling yours a) existing, and 
b) being included in any current force field, are geological. 
Accordingly, the odds against any current parameters providing an 
adequate physical model are cosmological.

> On the other hand I can use Gaussian software using the UFF forcefield. Because i am new to this field (master student doing an internship) I would like to know which is better to use, it's only to treat the whole protein plus clusters mechanically. 

Whoever set this task appears either to know little about the utility of 
MM forcefields, or to want to teach you about said utility :-)

> Later I will use Oniom methods with gaussian.

This is where you should start.

Mark